Symposium 2:

-Search and discovery of new cancer agents-

Abstract: Search and Discovery of New Cancer Agents for Breast Cancer

Lyndsay N. Harris, MD

Breast Cancer Program, Yale Medical Oncology, New Haven, USA

As we attempt to improve treatment of breast cancer it is critical that we better defi ne breast cancer subtypes to optimize the use of standard and targeted therapies. Newer technologies have led to the recognition of different breast tumor subtypes and these discoveries have improved our ability to defi ne optimal treatments for patients based on a better understanding of the biology of these subtypes. This overview will discuss the basic molecular defi nition of breast tumor subtypes and how this information is applied to the optimization of treatment for breast cancer patients.
The seminal work by Perou and Botstein, using microarray profi ling to defi ne the ‘intrinsic subtypes’, was the fi rst publication to describe a molecular classifi cation of breast cancer. Among the categories emerging from this study are the two categories of ER and/or PR positive (Luminal A and B) tumors, HER2 gene amplifi ed (HER2-enriched) tumors and ER/PR negative, HER2 negative (Basal-like) tumors. Despite differences in the platforms used for gene expression analysis, these subtypes can be identifi ed using microarray patterns from independent datasets. [1-4] While other categories of breast cancer exist (and will be added to the taxonomy) this classifi cation refl ects major subgroups of ductal breast cancer that are now used to subdivide patients when making treatment decisions.
From a therapeutic perspective, molecular classification is important, as it reduces the heterogeneity of patient groups and increases the likelihood of response to therapy. [5] There are two clear examples of this in breast cancer, from single-gene marker studies. The fi rst, and perhaps most important fi nding in the biology of breast cancer was the class distinction between ER positive and negative tumors. Across populations of breast cancer patients, it is clear that ER positive tumors respond to anti-estrogen therapy, while ER negative tumors do not [6-9]. A second example is that of HER2 gene amplifi ed tumors, which have been shown to respond preferentially to the anti-HER2 monoclonal antibody, trastuzumab (Herceptin).[10] It is gratifying, then, that expression profi ling is able to identify these subgroups, across platforms, and has further pointed out sets of genes that defi ne these tumor types. These subgroups are currently defi ned by cell lineage (luminal, basal) and the presence of the HER2 oncogene (HER2 enriched). The luminal tumors are divided into luminal A, and luminal B. While the initial study of Perou et al used several hundred genes to classify the tumor subtypes, this group has recently distilled the classifi cation to only 50 genes (termed the PAM50) [11]. In addition, they showed that the subtypes predict sensitivity to particular therapies. Furthermore, our recent analysis of a taxane monotherapy trial (CALGB 9842/9840) suggests that molecularly-defi ned tumor subtypes differ in their response to this commonly used chemotherapy agent. [12]

Responses of Breast Cancer Subtypes to Current Therapies
HER2-enriched Subtype

The HER2 tumor subtype is more sensitive to anthracycline-containing chemotherapy and may be less sensitive to alkylating agents and agents which produce DNA adducts such as Cisplatinum and Oxaliplatinum. [13-15] In addition, recent studies suggest that HER2 tumors are preferentially sensitive to taxanes. [16] There is evidence that HER2 targeted therapy can reverse drug resistance in this tumor subtype, which has led to the development of specifi c regimens which take advantage of this synergistic effect between HER2 antibodies and most chemotherapy regimens tested. [17, 18]

Luminal A and B Subtypes

Luminal tumors are driven, at least in part, by estrogen receptor signaling. However, it has long been appreciated that ER positive tumors vary in their response to anti-estrogens (eg. tamoxifen, aromatase inhibitors). One of the striking fi ndings from the molecular classifi cation of Luminal A and B tumors is that a set of genes (including ER) is overexpressed in ER positive tumors that are particularly sensitive to anti-estrogens and these genes are characteristic of the Luminal A subtype. Luminal B tumors express some of the genes in this ‘estrogen signature’ but typically have lower expression of ER, lack of expression of many signature genes and higher levels of proliferation genes. In addition, Luminal B tumors are more likely to express other growth factor receptors, are less sensitive to anti-estrogens and more sensitive to chemotherapy (8, 9). This information has profound implications for treatment decision-making and provides novel targets for this breast tumor subtype (see below).

Basal-like Subtype

A new subgroup of breast cancer, now termed the ‘basal-like’ or ‘triple-negative’ subtype, has emerged from microarray profi ling studies. This tumor type is molecularly distinct from other breast cancers, expressing one or more of the basal cytokeratins (CK 5/6, CK 14, CK903) and carries a worse outcome cancer, with up to 50% of these patients relapsing and dying of their breast cancer, even in early stages of disease. [1,4].
These breast cancers are insensitive to receptor-directed inhibitors (tamoxifen, anti-HER2 therapy), which may explain the worse prognosis associated with this tumor type across multiple breast cancer datasets. As noted above some basal tumors are likely to be taxane resistant. [19] However, the recently discovered ‘achilles heel’ of this tumor subtype may lie in the fact that these tumors appear to be defi cient in the homolgous recombination/repair (HSR) pathway, due to loss of BRCA1 activity. [20, 21] Indeed, the tumors from BRCA1 carriers are nearly always triple negative (basal-like) and profi le within the same group as sporadic basal tumors. The working hypothesis by many groups is that basal breast tumors are more sensitive to agents which induce DSBs such as cisplatin.

New Targets and Targeted Therapy for Breast Cancer Subtypes
HER2-enriched Subtype

The recognition of the pathogenic role of the HER2 oncogene in breast cancer led to the fi rst highly successful targeted therapy (trastuzumab, Herceptin®) that began with a scientifi c discovery (22, 10). This success story has demonstrated that biologic insights into the mechanism of breast cancer pathogenesis can lead to cures for our patients. In addition, the importance of this target has been proven time and again as newer HER2-targeted therapies continue to show high levels of activity in this breast cancer subtype (eg. lapatinib, pertuzumab, neratinib, TDM-1). Furthermore, therapies that target the pathways which are important in HER2 pathogenesis are also meeting with success (eg. HSP- 90 inhibitors, HDAC inhibitors). Finally, mechanisms of resistance to HER2 targeted therapy are predicted by known signal transduction pathways and alternative growth factor receptors. [23, 24, 25] The ‘HER2 paradigm’ demonstrates the success of the bench to bedside approach and gives hope and direction to scientists, clinicians and patients.

Luminal A and B Subtypes

While the ‘HER2 paradigm’ demonstrates the importance of the bench-to-bedside approach in developing targeted therapy, the fi rst targeted therapy in breast cancer are the anti-estrogens in ER/PR positive tumors, and reminds us of the importance of clinical observation. This class of drugs has been used for over 5 decades, based on the observation that a subset of breast tumors could be successfully treated by estrogen withdrawal using oophorectomy, adrenalectomy or hypophysectomy. It was not until the 1970’s that it was clearly appreciated what the target of these maneuvers was, ie the estrogen and/or progesterone receptors. The Luminal A and B subtypes represent the molecular phenotype that further refi nes this classifi cation, and provides important insights into variability in response to anti-estrogens in so-called ER/PR positive tumors. As noted above, the Luminal A subtype is highly sensitive to anti-estrogens with corresponding expression of genes which characterize the ‘estrogen signature’. The Luminal B tumors are less sensitive to anti-estrogens and have higher grade, proliferation and chemosensitivity. In addition, this subclass was discovered to have many other growth factor receptors (eg. c-MET, c-KIT, PDGFR, HER1-4) which provide numerous druggable targets. While the attempts to overcome anti-estrogen resistance with the addition of targeted therapy against these receptors has not been entirely successful, recent studies suggest that molecular profiling may be critical for understanding which tumors are most likely to benefi t from this approach. Specifi cally, co-targeted of ER and EGFR with tamoxifen and gefi tinib respectively was not a successful approach in the neoadjuvant setting [26], however more recent studies with lapatinib and letrozole suggest that the Luminal B subtype is most likely to benefi t from this approach [27]. In addition, a recent Phase II study showed highly promising activity of anastrazole combined with the multi-targeted kinase inhibitor sorafenib in aromatase-inhibitor resistant metastatic breast cancer [28].

Basal-like Subtype

Perhaps the most feared breast cancer subtype is the ‘basal-like’ tumor which lack the classic ER, PR and HER2 receptors, is virtually always poorly differentiated, angiogenic and shows a very poor outcome in older cohort studies without treatment. It is important to recognize, however, that many basal-like tumors are exquisitely sensitive to chemotherapy as evidenced by preoperative studies showing the highest pathologic complete response rate and excellent overall survival if pCR is achieved. [21] However, there is clearly a subgroup of basal-like tumors that have a poor prognosis despite chemotherapy and enormous efforts have been made to discover the ‘targets’ for this group. In fact, the biologic insights from molecular profi ling have played a critical role in the recent discovery and development of PARP inhibitors for basal-like breast cancer. As noted above, it was observed shortly after the defi nition of molecular subtypes that BRCA1 carriers almost always developed triple negative tumors and that these tumors profi led with the sporadic basal-like tumors. This led to the hypothesis that basal tumors were likely to be defi cient in homologous recombination (HSR) as this DNA repair defect is characteristic of BRCA1 defi ciency. Recent studies suggest that BRCA1 defi cient cells may be particularly sensitive to PARP inhibitors as these agents are able to disable the single-strand break repair pathway and this, coupled with the double-strand break repair defect in BRCA1 defi cient tumors leads to so-called synthetic lethality. [20, 29, 30] Although germ line mutations in BRCA1 account for the majority of dominantly inherited breast cancers, sporadic breast carcinomas rarely show mutations in the BRCA1 gene. [31] Interestingly, decreased BRCA1 expression has been observed in sporadic breast cancers correlating with higher tumor grade and poor prognosis, the basal-like subtype.[32,33] In this case, BRCA1 loss may be a result of epigenetic silencing of the BRCA1 promoter, or perhaps other mechanisms of BRCA downregulation, such as hypoxia which leads to decrease in BRCA1, RAD51 and other HSR pathway members.[34] Hence, the loss of BRCA1 appears to be associated with this tumor subtype, further supporting the contention that triple negative breast cancer is defi cient in HSR, a fact that has been exploited clinically with the use of cisplatin and PARP inhibitors.


Understanding of the basic biology and molecular profiles of breast cancer has led to important insights into the optimal treatment for patients. This has led to cures for patients suffering from this devastating disease and gives hope for the future development of new targeted therapy for breast cancer.

公益財団法人東京生化学研究会 CHAAO事業